rmw_send_reqponse returns RMW_RET_TIMEOUT.

[fujitatomoya]

part of https://github.com/ros2/ros2/issues/1253, but not dependent. this is doc fix for rmw_send_response to return RMW_RET_TIMEOUT.

• rmw_fastrtps

https://github.com/ros2/rmw_fastrtps/blob/926b3a1472fc5ba74972ca8384b054c94571b27c/rmw_fastrtps_shared_cpp/src/rmw_response.cpp#L148-L150

• rmw_cyclonedds

https://github.com/ros2/rmw_cyclonedds/blob/76572ebf1032f359ed302d7017471a295738fa1e/rmw_cyclonedds_cpp/src/rmw_node.cpp#L4541-L4542

• rmw_connnextdds does not return RMW_RET_TIMEOUT

[iuhilnehc-ynos]

Actually, I don't think it should return RMW_RET_TIMEOUT right now. Because there is no argument for users to call the rcl_send_response/rmw_send_response with a timeout value.

[iuhilnehc-ynos]

How about try-catch the rclcpp::exceptions::RCLError with a WARNING message for https://github.com/ros2/rclcpp/blob/beda0966db945ee3ba56e378183a7b7bfdab9526/rclcpp/src/rclcpp/executor.cpp#L581 similar to https://github.com/ros2/rclcpp/blob/beda0966db945ee3ba56e378183a7b7bfdab9526/rclcpp/src/rclcpp/executor.cpp#L571?

[fujitatomoya]

@iuhilnehc-ynos yeah i know, we need to think about the client behavior as well.

[fujitatomoya]

@clalancette @mjcarroll this is doc fix for rmw_send_reqponse, can you take a look at this?

[fujitatomoya]

CI:

• Linux
• Linux-aarch64
• Windows

[fujitatomoya]

CI (again, test option was wrong)

• Linux
• Linux-aarch64
• Windows

[fujitatomoya]

@clalancette @mjcarroll this is just doc fix aligns with implementation. there is nothing we do with RMW_RET_TIMEOUT for now, but rmw implementation does return RMW_RET_TIMEOUT.

[fujitatomoya]

This needs to be put back to humble. (https://github.com/ros2/rcl/pull/1048 has been merged to humble.)

[fujitatomoya]

@Mergifyio backport humble

[mergify[bot]]

backport humble

✅ Backports have been created

* [#367 rmw_send_reqponse returns RMW_RET_TIMEOUT. (backport #350)](https://github.com/ros2/rmw/pull/367) has been created for branch `humble`